• Keine Ergebnisse gefunden

The degradation of human recombinant tropoelastin by MMPs: mapping cleavage sites and

Fig. 3.3: Positive ion MALDI mass spectra of polypeptides generated when tropoelastin (shown by [M + H]+ in the chromatogram) was incubated for 30 min in the presence of MMP-7 (A, enzyme-to-substrate ratio 1:500); MMP-9 (B, 1:500); and MMP-12 (C, 1:5000).

3.2 The degradation of human recombinant tropoelastin by MMPs: mapping

during coacervation (Toonkool et al., 2001; Wise et al., 2005). It was, therefore, speculated that the incubation condition might support folding in the region and hide several bonds from MMPs attack. MMP-12 cleaved at least at 89 sites while MMP-7 and -9 cleaved at least at 58 and 63 sites, respectively. From the total cleavages, 24 were common sites for the three enzymes while 29 were shared by MMP-7 and MMP-9. Similarly, MMP-7 and MMP-12 shared 35 common cleavage sites, while MMP-9 and -12 exhibited the maximum common cleavage sites of 46. As also pointed out earlier, MMP-12 is the most active en-zyme against tropoelastin and it was identified to cut at several sites, except in the domains encoded by exons 8, 9, and 11. In contrast, MMP-7 and -9 did cleave multiple linkages in several domains but no cleavage could be detected in domains encoded by exons 8-11, 19, 20, and 36.

MMPs involve primarily their ctDs to facilitate cleavage of substrates. The interaction be-tween the zinc atom of the ctD and a residue constituting the scissile bond of the substrate has been identified to play the most important role in enzyme catalysis (Pirard, 2007).

Other residue-subsite contacts that take place at pockets either to the right or to the left of zinc have also been recognized to support cleavage processes. Subsites differ in the manner they bind to substrates; some bind strongly while others exhibit weak interactions. The difference in the binding affinities can be tapped to design unique inhibitors that are selec-tive in their interactions with a given MMP. However, the remarkable similarity between the 3D structure of MMPs and the resultant resemblance in substrate specificity always offer a formidable challenge to attain the objective of designing selective inhibitors. Fig.

3.4 depicts a ribbon structure of MMP-7, -9, and -12 aligned to demonstrate the degree of similarity these enzymes have.

A summary of specific residue-subsite interactions that show preferences of the respective subsites in the ctDs of the three MMPs is given in Tables 1 and 2 (appendix 6.2). As can be observed, the three MMPs cleaved typically N-terminal to hydrophobic and/or aliphatic residues such as Leu, Ala, Val, Gly, Ile, Tyr, and Phe with noticeable differences in their affinity towards individual residues. When residue preferences were determined on the basis of the number of cleavages each MMP initiated, the preference of the S1` in MMP-7 followed the order Leu (48 %) >> Val/Gly (each 12 %) > Pro (10 %) > Tyr (5 %). The residue specificity of MMP-9 followed the order Leu (22 %) > Ala (19 %) > Gly (14 %) >

Lys (11 %) > Val (9 %), while MMP-12 showed a specificity according to the order Ala (26 %) > Leu (20 %) > Lys (12 %) > Val/Tyr (each 9 %) > Gly (7 %). It is evident that the three enzymes showed certain degree of preferences for amino acids at S1`. To exemplify this, MMP-7 showed stronger preference for hydrophobic residues particularly for Leu (73

%, 29 out of a total of 40 Leu residues in tropoelastin) as compared to MMP-9 (35 %) and MMP-12 (48 %) at S1`. It is also interesting to note that Ala, a relatively small aliphatic and hydrophobic residue, was moderately preferred at S1` of MMP-9 and particularly pre-ferred in MMP-12. However, very few cleavages (2 out of 157 possible x-Ala bonds) could be identified with Ala at P1` for MMP-7. Moreover, all three MMPs appeared to cleave N-terminal to charged residues such as Lys with notable differences in their affinities. In par-ticular, around 31 % of all x-Lys linkages in tropoelastin could be cleaved by MMP-12, while in the case of MMP-7 and MMP-9 the proportions were 3 % and 20 %, respectively.

At other subsites, the three MMPs behaved more or less similarly and a detailed account has been provided in appendix 6.2.

Fig. 3.4: Alignment of the ribbon structures of MMP-7 (gray), MMP-9 (red), and MMP-12 (cyan).

Some of the observed residue-subsite interactions are interesting. Thus, to better explain these interactions the crystal structures of the three MMPs in complex with model peptides were graphically analyzed. As also described previously (Bode, 2003) and confirmed in this work (Fig. 3, appendix 6.2), the size and shape of individual subsites influenced resi-due-subsite interactions. For example, as evidenced by Fig. 3A, Tyr214 of the S1` loop in the

case of MMP7 was identified to extend into the opening of the loop, while MMP9 and -12 have the smaller Leu residue at the same position, giving the loop in MMP-9 and --12 an extended appearance. In addition, whereas the entrance of the S1` loop in MMP-9 was re-stricted by the bulkier but flexible Tyr393, the same position was occupied by the smaller Thr210 residue in MMP-12. Therefore, bulky and aromatic residues such as Ser, Lys, and Arg were restricted from binding in the shallower S1` pocket of MMP-7 but accommodated well in the deeper subsite of MMP-12. Furthermore, considering the hydrophobic nature of S1` charged residues are not expected to be favored. However, the degree of hydrophobic-ity differs between the three MMPs. For example, the polar residue Thr214 of MMP-12 is mutated to an aliphatic residue Val398 in MMP-9 and Ala215 in MMP-7. Glu219 of MMP-7 and MMP-12 is also mutated to Gln402 in MMP-9. Fig. 3C also confirmed that the bound model substrate with Lys at P1` rests in the vicinity, where it is more hydrophobic in MMP-7 than MMP-12 and MMP-9. Therefore, the more hydrophilic and wider pocket in MMP-12 could accommodate Lys better than MMP-7 and -9. The other interesting resi-due-subsite interaction is that Pro was a preferred residue at P3 (Tables 1 and 2, appendix 6.2). Modeling of the crystal structure of MMP-12 in complex with the natural substrate LPYGYGP (residues 226–233, tropoelastin isoform 2) revealed that Pro at P3 appeared to force the backbone of the peptide into a conformation that favors the occupation of indi-vidual subsites (Fig. 4A, appendix 6.2). Furthermore, the positioning of Pro at P3 allowed bulky residues such as Tyr at P2 position to interact with S2 (Figs. 4B and C, appendix 6.2).

Overall, as discussed in section 3.1 tropoelastin is degraded by MMPs rapidly and it is now known that most cleavages take place mainly close to the two terminal regions of the tro-poelastin sequence. The rapid cleavages in the highly conserved C-terminal region, in par-ticular, are interesting. The C-terminal domain of tropoelastin has been proven to be re-sponsible for tropoelastin’s binding with microfibrillar proteins and possibly cross-linking during elastin biosynthesis (Broekelmann et al., 2008; Sato et al., 2007). Products of ter-minally truncated tropoelastin sequences have been shown to cause errors in elastin bio-synthesis. Examples of similar cases include SVAS and lamb ductus arteriosis diseases (Hinek and Rabinovitch, 1993; Tassabehji et al., 1997; Wu and Weiss, 1999). Thus, the present study identified the susceptibility of tropoelastin to MMPs and further mapped sus-ceptible bonds. The clear indication of these results is that tropoelastin will be rapidly processed when it is expressed in the presence of MMPs. Moreover, as previous results

showed, processed tropoelastin cannot effectively participate in elastin synthesis. There-fore, this underlines not only the direct degradative roles of MMPs on tropoelastin but it also shows the potential influence MMPs have on the ECM functions by undermining new synthesis and repair of damaged elastin. The information gained by mapping susceptible bonds in tropoelastin can also be applied in the designing of biomaterials based on tropoe-lastin. Efforts have already been made to develop resistant tropoelastin derivatives using recombinant technologies or chemical methods by modifying susceptible regions of tro-poelastin without affecting functional properties (Ng et al., 2008; Weiss, 2007). Further-more, the cleavage site specificity study has provided several important results that can be used to understand tropoelastin-MMP interactions.